What is a Claude competitor research skill?

Quick Answer: A Claude competitor research skill is a reusable SKILL.md instruction file that turns Claude into a structured competitive intelligence agent. You define the intake (what company to research, which dimensions to cover) and the output format (battlecard, spreadsheet, summary doc) once, then run deep competitor research on demand without rewriting prompts every time.

Most competitor research in Claude goes the same way. You type a vague prompt, get a surface-level summary of a company's homepage, and walk away with nothing you couldn't have found in 10 minutes on Google. The problem is not Claude. The problem is the prompt.

Claude Skills fix this. By building a dedicated competitor research skill with a defined structure, intake form, and output template, you get consistent, repeatable intelligence you can actually use in positioning, sales, and product decisions. If you need outside support turning that intelligence into pipeline, it can also help to benchmark against specialist B2B SaaS SEO agencies or broader B2B SaaS digital marketing agencies.

This guide walks you through exactly how to build that skill, what to put in it, and what good output looks like.

What Is a Claude Skill?

A Claude Skill is a saved, reusable instruction set that tells Claude how to behave for a specific recurring task. Think of it as a standing operating procedure written in plain language that Claude reads before it starts any work.

Skills live in Claude's Projects feature as a file (commonly called SKILL.md or a system prompt block). Every time you run the skill, Claude follows the same logic, asks for the same inputs, and produces output in the same format.

For competitor research, this matters because consistency is the whole point. You want the same depth of analysis whether you are researching a direct rival or a new entrant you spotted at a conference.

Why Most Claude Competitor Research Fails

Before building the skill, understand where ad-hoc research breaks down:

• No scope definition. Claude does not know if you want a 3-line summary or a 2,000-word battlecard unless you tell it.
• No research dimensions. Without structure, Claude defaults to the obvious: homepage copy, product description, pricing page. It skips positioning signals, ICP clues, and messaging gaps.
• No output format. Every result looks different, which makes comparison across multiple competitors almost impossible.
• No intake. You re-explain the context every single time.

A well-built SKILL.md solves all four problems at once.

The Anatomy of a Competitor Research SKILL.md

A strong competitor research skill has four parts: a role definition, an intake form, a research framework, and an output template.

Part 1: Role Definition

This tells Claude what it is doing and why. Keep it tight. One short paragraph is enough.

You are a B2B SaaS competitive intelligence analyst. Your job is to produce 
structured, evidence-based competitor profiles that a product marketer or 
sales team can use immediately. You do not speculate. You cite observable 
signals: website copy, pricing pages, job listings, review sites, and 
published content. When you cannot find a data point, you say so.

The "you do not speculate" line is doing real work here. Without it, Claude fills gaps with plausible-sounding guesses that look like facts.

Part 2: The Intake Form

This is what you fill in each time you run the skill. It replaces the "here's what I need" paragraph you would otherwise write from scratch.

## Intake

- **Competitor name:** [Name]
- **Competitor URL:** [URL]
- **Our product:** [Your product name and one-line description]
- **Our ICP:** [Describe your target customer: company size, industry, role]
- **Research depth:** [Quick scan (15 min) / Standard (30 min) / Deep dive (60 min)]
- **Output format:** [Battlecard / Full profile / Positioning gap analysis]
- **Specific questions to answer:** [Optional: list any specific angles to investigate]

The research depth field is optional but useful. It tells Claude how much to expand on each section rather than defaulting to maximum verbosity on every run.

Part 3: The Research Framework

This is the core of the skill. It defines exactly which dimensions Claude must investigate and what signals to look for in each one.

## Research Framework

Work through each dimension below in order. For each one, pull observable 
evidence before drawing conclusions.

### 1. Positioning and Messaging
- What problem do they say they solve? Quote their headline and subheadline.
- Who do they explicitly name as their customer?
- What outcome do they promise?
- What words and phrases repeat across their homepage, about page, and ads?

### 2. Product and Features
- What is the core product?
- What integrations do they highlight?
- What features appear in their navigation, pricing tiers, or case studies?
- What do they NOT mention that we would expect them to cover?

### 3. Pricing and Packaging
- Is pricing public? If yes, list tiers, prices, and what each includes.
- What is the entry point (free trial, freemium, demo-only)?
- What limits or gates appear at each tier?

### 4. ICP Signals
- What company sizes, industries, or roles appear in their case studies?
- What logos do they display?
- What job titles appear in testimonials?
- What verticals do their blog posts target?

### 5. Go-to-Market Motion
- Do they appear to be product-led, sales-led, or a hybrid?
- What content do they publish and at what frequency?
- What channels are they active on (LinkedIn, G2, review sites, ads)?
- Do they have a partner or reseller programme?

### 6. Weaknesses and Gaps
- What do G2, Capterra, or Trustpilot reviews complain about?
- What features are conspicuously absent from their positioning?
- Where does their messaging feel vague or defensive?

### 7. Competitive Angle Against Us
- Where do they overlap directly with our product?
- Where do they serve a different segment?
- What is the strongest objection a prospect might raise after seeing them?
- What is our clearest differentiator against this competitor?

Each dimension produces a section of the final output. The "Weaknesses and Gaps" section is the one most ad-hoc research skips entirely. It is also the most useful for sales teams, especially when paired with a clear positioning and content marketing strategy.

Part 4: The Output Template

Define exactly what the finished document looks like. This makes outputs comparable across multiple competitors.

## Output Format: Battlecard

Produce the output in the following structure. Use headers exactly as written.

---

# Competitor Battlecard: [Competitor Name]
**Researched:** [Date]
**Depth:** [Quick / Standard / Deep]

## One-Line Summary
[A single sentence describing who they are and who they serve]

## Positioning
- **Their headline:** [Direct quote]
- **Core promise:** [What outcome they sell]
- **Target customer:** [As stated or implied by their site]

## Product Snapshot
[3-5 bullet points on core features and differentiators]

## Pricing
[Tier structure or "pricing not public" with entry point noted]

## ICP Evidence
[2-4 bullet points on company types, sizes, and roles visible in their content]

## GTM Motion
[2-3 sentences on their primary growth channel and content approach]

## Their Weaknesses
[3-5 bullet points drawn from reviews or positioning gaps]

## How to Win Against Them
- **Lead with:** [Your strongest differentiator]
- **Handle the objection:** [Most likely objection and how to counter it]
- **Avoid:** [Any area where comparison does not favour us]

---

How to Install the Skill in Claude

1. Open Claude and go to Projects.
2. Create a new project named something like "Competitive Intelligence".
3. In the project settings, paste your full SKILL.md content into the Custom Instructions or Project Instructions field.
4. Save the project.

Every conversation you start inside this project now runs with the skill active. To research a competitor, open a new conversation, paste the completed intake form, and type "Run the competitor research skill."

Running Your First Competitor Research Session

Once the skill is installed, a research session takes three steps.

Step 1: Fill in the intake form

Copy the intake block from Part 2, fill in the fields for your target competitor, and paste it into the conversation.

Step 2: Give Claude access to sources

For best results, paste in raw text from the competitor's homepage, pricing page, and one or two case studies. Claude's web browsing (if available in your plan) can supplement this, but pasted content gives you more control over what it reads.

For review site data, copy 10-15 recent reviews from G2 or Capterra and paste them under a heading like "Review data: [Competitor name]".

Step 3: Run the skill

Type: "Run the competitor research skill using the intake above and the source material provided."

Claude will work through each dimension in the framework and produce the battlecard in the exact format defined in the template.

What Good Output Looks Like

A well-run competitor research skill produces output that passes this test: hand it to a sales rep with no context and they should be able to use it in a discovery call the same day.

Specifically, good output includes:

• Direct quotes from the competitor's site, not paraphrases
• Specific logos or customer names from their case studies, not "enterprise customers"
• Actual pricing numbers or a clear note that pricing is gated
• Review-sourced weaknesses with the source noted (G2, Capterra, etc.)
• A concrete differentiator written in the language your sales team actually uses

If the output reads like a Wikipedia summary, the research framework is not specific enough. Add more signal-seeking instructions to Part 3.

Extending the Skill for Deeper Research

Once the base skill works, three extensions are worth adding.

Multi-competitor comparison mode. Add an output variant that produces a side-by-side comparison table across 3-5 competitors. Define the columns (positioning, pricing, ICP, GTM motion) and ask Claude to populate each cell from the individual battlecards you have already created.

Job listing analysis. Competitor job listings reveal roadmap priorities before any product announcement. Add a dimension to the framework: "If job listings are provided, identify which functions they are hiring into and what this suggests about their product direction."

Content gap analysis. Paste in 10-15 of their recent blog post titles. Add a framework dimension: "Identify which topics they publish consistently, which topics they avoid, and where our content could credibly own a position they have left open." If that analysis informs campaign planning, it often overlaps with work handled by B2B SaaS inbound marketing agencies or specialist B2B SaaS ABM agencies, depending on your go-to-market model.

FAQs

What is a Claude competitor research skill? A Claude competitor research skill is a saved SKILL.md instruction file that gives Claude a defined role, intake form, research framework, and output template for competitive analysis. It replaces ad-hoc prompting with a repeatable process that produces consistent, structured battlecards every time you run it.

How is a Claude skill different from a regular prompt? A regular prompt is written fresh each time and produces inconsistent results. A Claude skill is a persistent instruction set stored in a Project that Claude reads before every session. The skill defines scope, method, and format once so you never have to re-explain the task.

Can Claude do competitor research without browsing the web? Yes. The most reliable approach is to paste source material directly into the conversation: homepage copy, pricing pages, case studies, and review site text. Claude analyses what you give it. If your Claude plan includes web browsing, it can supplement this, but pasted content gives you more control over source quality.

Is SaaS Hackers' SKILL.md template free to use? The template in this guide is free to copy and adapt. Paste it into your Claude Project instructions, adjust the intake fields and output format to match your workflow, and it is ready to run.

How often should I update a competitor battlecard? For active competitors in a fast-moving market, a quarterly refresh is the minimum. Set a recurring task to re-run the skill every 90 days or immediately after a competitor announces a pricing change, new feature, or funding round.